Commits

Breadcrumbs

History for

onbe3b16d

User selector

All users

Datepicker

All time

Commit History

Commits on Jan 5, 2023

• Apply clang-tidy perf improvements to aten and torch/jit/passes/onnx (pytorch#91726)

  Show description for 3d17728
  

  Skylion007

  authored and

  pytorchmergebot

  committed
  3d17728

  Copy full SHA for 3d17728

Commits on Oct 9, 2022

• Quantized: Use per-operator headers (pytorch#75569)

  Show description for 7134b9b
  

  peterbell10

  authored and

  pytorchmergebot

  committed
  7134b9b

  Copy full SHA for 7134b9b

Commits on Sep 30, 2022

• Create a quantized in-palce version CUDA ReLU function, relu_quantized_cuda_. (pytorch#85670)

  Show description for 0a7d8b4
  

  fufeisi

  authored and

  pytorchmergebot

  committed
  0a7d8b4

  Copy full SHA for 0a7d8b4

Commits on Sep 29, 2022

• Create a quantized non-in-palce version CUDA ReLU function, (pytorch#85669)

  Show description for 3c9e8cd
  

  fufeisi

  authored and

  pytorchmergebot

  committed
  3c9e8cd

  Copy full SHA for 3c9e8cd

Commits on Sep 24, 2022

• Revert "Create a quantized version ReLU function for CUDA (pytorch#85502)"

  Show description for 41be45f
  pytorchmergebot

  committed
  41be45f

  Copy full SHA for 41be45f

• Create a quantized version ReLU function for CUDA (pytorch#85502)

  Show description for 93a53ff
  

  fufeisi

  authored and

  pytorchmergebot

  committed
  93a53ff

  Copy full SHA for 93a53ff

Commits on Jun 7, 2022

• [quant][core][better-engineering] Rename files in quantized directory to conform with non-quantized countertpart filenames

  Show description for a56f4e2
  

  dzdang

  authored and

  pytorchmergebot

  committed
  a56f4e2

  Copy full SHA for a56f4e2

Commits on May 25, 2022

• [quant][core][better-engineering] Rename files in quantized/cuda directory to conform with non-quantized countertpart filenames pytorch#77431

  Show description for a9b6800
  

  dzdang

  authored and

  pytorchmergebot

  committed
  a9b6800

  Copy full SHA for a9b6800

• [quant][core][improvement][bug fix] Added channel axis bound checking in fused_moving_avg_obs_fake_quant_*

  Show description for ba952d8
  

  dzdang

  authored and

  pytorchmergebot

  committed
  ba952d8

  Copy full SHA for ba952d8

Commits on May 24, 2022

• [quant][core][gpu][feature] Implemented quantized cuda gelu

  Show description for 2aad28a
  

  dzdang

  authored and

  pytorchmergebot

  committed
  2aad28a

  Copy full SHA for 2aad28a

Commits on May 14, 2022

• Revert "[quant][core][gpu][feature] Implemented quantized cuda gelu"

  Show description for 981719f
  pytorchmergebot

  committed
  981719f

  Copy full SHA for 981719f

Commits on May 13, 2022

• [quant][core][gpu][feature] Implemented quantized cuda gelu

  Show description for b892b85
  

  dzdang

  authored and

  pytorchmergebot

  committed
  b892b85

  Copy full SHA for b892b85

Commits on Oct 8, 2021

• Remove cuda/Loops.cuh dependency on native_functions.yaml (pytorch#64168)

  Show description for 731cf49
  

  peterbell10

  authored and

  facebook-github-bot

  committed
  731cf49

  Copy full SHA for 731cf49

Commits on Oct 6, 2021

• [quant] Add fp32/fp16 zero_point support for GPU fakeQuant (pytorch#65836)

  Show description for 83bac89
  

  alexanderguzhva

  authored and

  facebook-github-bot

  committed
  83bac89

  Copy full SHA for 83bac89

Commits on Oct 5, 2021

• Back out "[quant] update fused_obs_fake_quant op to accept output_fake_quant argument" (pytorch#66063)

  Show description for 458a00b
  

  supriyar

  authored and

  facebook-github-bot

  committed
  458a00b

  Copy full SHA for 458a00b

Commits on Sep 30, 2021

• [quant] update fused_obs_fake_quant op to accept output_fake_quant argument (pytorch#65621)

  Show description for 4666e3f
  

  supriyar

  authored and

  facebook-github-bot

  committed
  4666e3f

  Copy full SHA for 4666e3f

Commits on Sep 28, 2021

• Migrate remainder of THCDeviceUtils.cuh to ATen (pytorch#65472)

  Show description for 46b3fc0
  

  peterbell10

  authored and

  facebook-github-bot

  committed
  46b3fc0

  Copy full SHA for 46b3fc0

Commits on Sep 13, 2021

• [Pytorch Edge] Quantized Ops Dtype Selective (pytorch#63680)

  Show description for fd716fc
  

  JacobSzwejbka

  authored and

  facebook-github-bot

  committed
  fd716fc

  Copy full SHA for fd716fc

Commits on Aug 27, 2021

• [pytorch] add per_sample_weights support for embedding_bag_4bit_rowwise_offsets (pytorch#63605)

  Show description for 0c9dce9
  

  another-pjohnson

  authored and

  facebook-github-bot

  committed
  0c9dce9

  Copy full SHA for 0c9dce9

Commits on Aug 12, 2021

• [quant] Make version 1 the default for get_default_qat_qconfig (pytorch#63043)

  Show description for d5a7579
  

  supriyar

  authored and

  facebook-github-bot

  committed
  d5a7579

  Copy full SHA for d5a7579

Commits on Aug 2, 2021

• [quant] Support PerChannel quantization in FusedMovingAvgObsFakeQuantize (pytorch#62346)

  Show description for aa5e3ad
  

  supriyar

  authored and

  facebook-github-bot

  committed
  aa5e3ad

  Copy full SHA for aa5e3ad

Commits on Jul 26, 2021

• Add some missing cuda guards (pytorch#62100)

  Show description for f9dce59
  

  r-barnes

  authored and

  facebook-github-bot

  committed
  f9dce59

  Copy full SHA for f9dce59

Commits on Jul 21, 2021

• [pytorch] Support embedding_bag_4bit_rowwise_offsets in cuda (pytorch#61728)

  Show description for f41d334
  

  another-pjohnson

  authored and

  facebook-github-bot

  committed
  f41d334

  Copy full SHA for f41d334

• [quant] Remove calls to .item() for fake_quant_on (pytorch#61921)

  Show description for ea8abcf
  

  supriyar

  authored and

  facebook-github-bot

  committed
  ea8abcf

  Copy full SHA for ea8abcf

• [quant] Add a new fused MovingAvg Obs + FakeQuant operator (GPU) (pytorch#61589)

  Show description for afdca41
  

  supriyar

  authored and

  facebook-github-bot

  committed
  afdca41

  Copy full SHA for afdca41

Commits on Jul 11, 2021

• [quant] Add tensor_qparam variant to fake_quantize_per_tensor (pytorch#61317)

  Show description for 99848c7
  

  supriyar

  authored and

  facebook-github-bot

  committed
  99848c7

  Copy full SHA for 99848c7

Commits on Jul 9, 2021

• update internal function names that apply to both cpu and cuda (pytorch#59701)

  Show description for 06dfaad
  

  HDCharles

  authored and

  facebook-github-bot

  committed
  06dfaad

  Copy full SHA for 06dfaad

Commits on Jul 8, 2021

• [bootcamp][pytorch][WIP] Support embedding_bag_byte_rowwise_offsets in cuda (pytorch#61075)

  Show description for c76eba6
  

  another-pjohnson

  authored and

  facebook-github-bot

  committed
  c76eba6

  Copy full SHA for c76eba6

• Cuda quantized tensors, support for quantize per channel (pytorch#58245)

  Show description for d51b437
  

  HDCharles

  authored and

  facebook-github-bot

  committed
  d51b437

  Copy full SHA for d51b437

Commits on Jul 1, 2021

• add embedding bag skeleton take 2 (pytorch#61126)

  Show description for d2fef35
  

  Hector Yuen

  authored and

  facebook-github-bot

  committed
  d2fef35

  Copy full SHA for d2fef35

• Revert D29313058: add an embedding bag skeleton operators

  Show description for 7017dc1
  

  Rong Rong (AI Infra)

  authored and

  facebook-github-bot

  committed
  7017dc1

  Copy full SHA for 7017dc1

• add an embedding bag skeleton operators (pytorch#60491)

  Show description for ae21357
  

  Hector Yuen

  authored and

  facebook-github-bot

  committed
  ae21357

  Copy full SHA for ae21357

Commits on May 26, 2021

• [PyTorch] Rename TI::add_borrowed_{in,out}put to TI::add_{in,out}put (pytorch#58608)

  Show description for ac67cda
  

  swolchok

  authored and

  facebook-github-bot

  committed
  ac67cda

  Copy full SHA for ac67cda

Commits on May 19, 2021

• [ATen][Quant] Pass at::Tensor by reference (pytorch#58284)

  Show description for 954d39b
  

  Hao Lu

  authored and

  facebook-github-bot

  committed
  954d39b

  Copy full SHA for 954d39b

Commits on May 18, 2021

• [PyTorch] Migrate remaining CUDA TI usage to borrowing where possible (pytorch#58278)

  Show description for 03f2f0f
  

  swolchok

  authored and

  facebook-github-bot

  committed
  03f2f0f

  Copy full SHA for 03f2f0f

Pagination

Previous

Next